Red Hat Bugzilla – Bug 217798
Need to port Resource Group optimization from gfs1 to gfs2
Last modified: 2010-01-11 22:36:52 EST
Description of problem:
For bz213763, we changed gfs_mkfs so that it can optimize the size
of the Resource Groups (RGs) for best performance. We need to port
that fix from gfs1 (gfs_fsck) to gfs2 (mkfs.gfs2).
Version-Release number of selected component (if applicable):
Steps to Reproduce:
mkfs.gfs2 -t marathon:gfs_vs0 -p lock_dlm -j 5 /dev/linear_1/linear_10
where /dev/linear_1/linear-10 is a 40TB logical volume.
mkfs.gfs2 uses the old default, 256MB RG size, which requires
more than 150000 RGs.
mkfs.gfs2 should figure out that 256MB would require more than 10000
RGs, and therefore it should use 2048MB RGs for best performance.
Prototype patch is already coded.
Created attachment 142454 [details]
Patch to fix the problem
Unit tested on marathon-01.
Made the title more descriptive.
This is a port of the fix for bug 213763.
I committed the fix to CVS in the HEAD, RHEL5 and RHEL50 branches.
A package has been built which should help the problem described in
this bug report. This report is therefore being closed with a resolution
of CURRENTRELEASE. You may reopen this bug report if the solution does
not work for you.